pythia-pgs floating point number precision issue for lepton colliders

[Sho Iwamoto]

Hi guys,

I am escalating this question to a bug of pythia6.
https://answers.launchpad.net/mg5amcnlo/+question/644347/

Now, by switching e<->p, the unweighted_events.lhe.gz stores events with "electron-energy == beam energy" for e-p collision

7 1 +1.0337370e-05 3.06904600e+02 7.81763000e-03 1.20735500e-01
       11 -1 0 0 0 0 +0.0000000000e+00 +0.0000000000e+00 +6.0000000000e+01 6.0000000000e+01 0.0000000000e+00 0.0000e+00 -1.0000e+00

However, pythia-pgs fails in pyremn.f (Line 323):

      IF(PMIN(0)+PMIN(1)+PMIN(2).GT.VINT(1).OR.(MINT(45).GE.2.AND.
     &PMIN(1).GT.PSYS(1,4)).OR.(MINT(46).GE.2.AND.PMIN(2).GT.
     &PSYS(2,4))) THEN

because MINT(45) == 3 (e-beam), PMIN(1) == 4.7e-16 and PSYS(1,4) == 0
and vetoes lots of the events saying

      Error type 9 has occured after 0 PYEXEC calls:
      (PYEVNW:) failed to evolve shower or multiple interactions. Returning.

This is because pythia6 does not care the floating number point precision seriously.
I am not sure you want to fix this issue (because pythia6 is no longer maintained) but anyway am reporting this issue to share with others.

[Sho Iwamoto (misho)]

maybe: https://answers.launchpad.net/mg5amcnlo/+question/268194 ?

[Sho Iwamoto (misho)]

A dirty patch works for me, but really ugly.

--- a/libraries/pylib/src/pyremn.f
+++ b/libraries/pylib/src/pyremn.f
@@ -321,8 +321,8 @@ C...Check minimum invariant mass of remnant system(s).
      & P(MINT(83)+JT+2,2)**2)
   340 CONTINUE
       IF(PMIN(0)+PMIN(1)+PMIN(2).GT.VINT(1).OR.(MINT(45).GE.2.AND.
- &PMIN(1).GT.PSYS(1,4)).OR.(MINT(46).GE.2.AND.PMIN(2).GT.
- &PSYS(2,4))) THEN
+ &PMIN(1).GT.PSYS(1,4)+1d-8).OR.(MINT(46).GE.2.AND.PMIN(2).GT.
+ &PSYS(2,4)+1d-8)) THEN
         MINT(51)=1
         MINT(57)=MINT(57)+1
         RETURN